JP6501900B2 - Apparatus and method for controlling a robot manipulator - Google Patents

Description

The present invention controls a robot manipulator comprising an end effector and driven by a plurality (M) of actuators AKT _m (where m = 1, 2, ..., M). Apparatus and method The invention further relates to a computer system equipped with a data processing device, a digital storage medium, a computer program product and a computer program.

  Robot manipulators are known to be superior to humans in terms of speed repeatability and accuracy when performing operations. However, when the application force has to be adjusted delicately, or when compliance has to be taken into consideration, humans are still better than robot manipulators, which is especially important in It is clear from the practical application example in which the operation operation and the assembly operation of the object become delicate operations. In particular, in delicate assembly work, the coordination between the magnitude of the force applied to the contact portion and the movement path is complicated.

  In this regard, it is known to apply "impedance control" to robot manipulators. The technical concept of applying impedance control to a robot manipulator is aimed at applying an active control to the robot manipulator so that the robot mimics the task being performed by a human being. The active control is performed by using, for example, a model consisting of a mass element, a spring element and a damping element to control the model from the outside.

  Generally speaking, in order to provide the robot manipulator with the desired compliance, the robot manipulator may be provided with active control, or components having compliance may be incorporated into the robot manipulator, or they may be I try to use together. However, it is also known that compliance in an arbitrary coordinate direction can not be imparted only by setting a joint of a robot / manipulator as a joint capable of unconstrained elastic rotational deviation (Non-patent Document 1). (See Albu-Schaffer, Fischer, Schreiber, Schoeppe, & Hirzinger (2004)), so it is not only necessary to provide robot manipulators with passive compliance, but also in addition to active control. Yes, it is not possible to avoid the occurrence of a problem Only by combining the active control, the application of the specified size to the surrounding environment without being affected by the inaccuracies of the operation object model or the surface model It is possible to apply a force and / or perform an operation as specified on the operation object.

  Furthermore, there is "active interaction control" as a known technique. Among the active interaction control, there are "direct" application force control and "indirect" application force control (see Non-patent document 14 (see Villani & De Schutter (2008).). The applied force control is used in combination with various virtual position techniques (see Non-patent Document 11 (Lutscher & Cheng (2014) and Non-patent Document 10 (Lee & Huang (2010)), and others). There are known techniques of applied force control, position control, and / or impedance control which are performed under given constraint conditions in various spaces (Borghesan & De Schutter 2014)).

  Although the progress in the field of control technology of robot manipulators is remarkable, the following drawbacks still exist so far.

  First, in order to generate the target application force to the robot manipulator to which only the impedance control is applied, it is performed by completely feedforward control, or the target of the robot manipulator end effector Either the use of a virtual target position where a shift has been introduced to the position has been performed. In these control methods, the magnitude of the external force is not clearly understood, but the target application force with sufficient accuracy from the end effector of the robot manipulator to the surrounding environment / operation target / workpiece etc. / In order to apply the target applied torque, it is necessary to clearly grasp the magnitude of the external force. Furthermore, in order for these control schemes to function, the surrounding environment also needs to be modeled with sufficient accuracy in terms of its geometry and compliance characteristics. However, this is contrary to the basic idea of impedance control to be able to function even if the surrounding environment is not modeled.

  Furthermore, as a disadvantage associated with robot manipulators to which only impedance control is applied, an application force of a target magnitude by feedforward control is manipulated from the end effector of the robot manipulator ( When acting on the ambient environment, if the contact between the end effector and the operation object disappears, the robot manipulator performs a large movement (ie, movement of the end effector) at that moment. Paths, moving speeds, and moving accelerations can produce large movements), and such movements carry potential risks. In addition, such movement occurs, for example, when the virtual target position of the end effector is set at a position far away from the actual position of the end effector at the moment when the contact state disappears. It is a cause.

  Furthermore, it is also known to "apply only applied force control" to a robot manipulator. The basic idea of applying the applied force control to the robot manipulator is to operate or manipulate the operation object by making the magnitude of the external force acting between the robot and the surrounding environment sufficiently accurate. It is intended to be able to manipulate the surface of the object with high accuracy. Providing this capability is one of the key requirements for using robotic manipulators in industrial applications. Therefore, the low precision impedance control and the like conventionally used can not replace the applied force control. As a solution to this problem, a so-called hybrid position application force control method has been proposed, and there are many types of this control method (see Non-Patent Document 12 (Raibert & Craig (1981)). The basic idea of this hybrid position application force control method is to divide a so-called "task space" into a plurality of complementary application position spaces, and apply an application force or an application torque in each of the plurality of spaces. It is to make it act and control the movement.

  A disadvantage associated with the various known hybrid force control schemes is that they have very low robustness with regard to loss of contact between the robot manipulator and the surrounding environment. Furthermore, such a control scheme can not ensure good control performance unless the surrounding environment is modeled with very high accuracy, but it is rare to obtain a model with sufficiently good accuracy. I can not.

  In order to clarify the stability of this type of control device, the model when modeling the surrounding environment is usually in the form of a simple system consisting of a spring element and a damping element. Further, Non-Patent Document 15 (Zeng & Hemami (1997)) is a document that comprehensively discusses various applied force control devices, including even a reference to the stability analysis of the control devices. In addition, general criticism on this kind of control device is made in non-patent document 5 (Duffy (1990)), and as the basis of the criticism there, the measurement method and the coordinate system display method can not be appropriately selected. It is mentioned that there are many things.

Albu-Schaffer, A., Fischer, M., Schreiber, G., Schoeppe, F., & Hirzinger, G. (2004). Soft robotics: What Cartesian stiffness can obtain with passively compliant, uncoupled joints. IROS. Albu-Schaffer, A., Ott, C., & Hirzinger, G. (2007). A Unified Passivity-based Control Framework for Position, Torque and Impedance Control of Flexible Joint Robots. The Int. J. of Robotics Research , (S. 23-39). Borghesan, G., & De Schutter, J. (2014). Constraint-based specification of hybrid position-impedance-force tasks. IEEE International Conference on Robotics and Automation 2014 (ICRA 2014). Cervera, J., Van Der Schaft, A., & Banos, A. (2007). Interconnection of port-Hamiltonian systems and composition of Dirac structures. Automatica (S. 212-225). Elsevier. Duffy, J. (1990). The fallacy of modern hybrid control theory that is based on orthogonal complements of twist and wrench spaces. (S. 139-144). Wiley Online Library. Duindam, V., & Stramigioli, S. (2004). Port-based asymptotic curve tracking for mechanical systems. (S. 411-420). Elsevier. Haddadin, S. (2013). Towards Safe Robots: Approaching Asimov's 1st Law. Springer Publishing Company, Incorporated. Haddadin, S., Albu-Schaffer, A., De Luca, A., & Hirzinger, G. (2008). Collision detection and reaction: A contribution to safe physical human-robot interaction. Intelligent Robots and Systems (S) 3356-3363). Hogan, N. (1985). [Impedance Control: An approach to manipulation: Part I-Theory, Part II-Implementation, Part III-Applications. ASME Journal of Dynamic Systems, Measurement, and Control, (S. 1-1 twenty four). Lee, D., & Huang, K. (2010). Passive-set-position-modulation framework for interactive robotic systems. IEEE Transactions on Robotics (S. 354-369). Lutscher, E., & Cheng, G. (2014). Constrained Manipulation in Unstructured Environments Utilizing Hierarchical Task Specification for Indirect Force Controlled Robots. IEEE International Conference on Robotics and Automation 2014 (ICRA 2014). Raibert, M. H., & Craig, J. J. (1981). Hybrid position / force control of manipulators. ASME Journal of Dynamical Systems, Measurement and Control, (S. 126-133). Spong, M. (1987). Modeling and Control of Elastic Joint Robots. ASME J. on Dynamic Systems, Measurement, and Control, (S. 310-319). Villani, L., & De Schutter, J. (2008). Force Control. In O. Khatib, Springer Handbook of Robotics (S. 161-185). Zeng, G., & Hemami, A. (1997). An overview of robot force control. Robotica (S. 473-482). Cambridge Univ Press.

The object of the present invention is a robot manipulator comprising an end effector and driven by a plurality (M) of actuators AKT _m (where m = 1, 2,..., M). It is an object of the present invention to provide an apparatus and method for controlling a number of the above-mentioned disadvantages in the apparatus and method. In particular, the object is to prevent the occurrence of the large movement of the robot manipulator which has occurred when the contact between the end effector and the operation object disappears.

  The invention consists of the components described in the independent claims. The dependent claims are based on a particularly advantageous embodiment. Further features, applications and advantages of the present invention will be apparent from the following description and in particular from the description of the embodiments of the present invention shown in the drawings.

According to a first aspect of the present invention, the above object is provided with an end effector and driven by a plurality (M) of actuators AKT _m (where m = 1, 2, ..., M) This is achieved by an apparatus for controlling the robot manipulator. Here, the term "actuator" is used in a broad sense, and this term includes, for example, electric motors, hydraulic motors, linear motors, step motors, and piezo actuators. And also others.

The apparatus further detects and / or detects a craft vinder F → _ext (t) = [f → _ext (t), m → _ext (t)] representing the external force / external torque acting on the end effector. A first unit to be derived is provided, where f → _ext (t) represents an external force acting on the end effector, and m → _ext (t) represents an external torque acting on the end effector. In the present disclosure, an arrow representing a vector is usually attached right above the character to indicate the right shoulder of the character. Therefore, “F → _ext (t)”, “f → _ext (t) ", the notation" m → _ext (t) "is they represent that the vector function of time. In the following description, not only arrows but also dots, bars, and tildes that are usually placed just above letters are placed on the right shoulder of letters. The first unit may be a craft vinder F → _ext (t) = [f → _ext (t), m → _ext (t) representing the external force / external torque acting on the end effector to provide its function. A sensor system for detecting a) and / or a craft windower representing the external force / external torque acting on the end effector-F → _ext (t) = [f → _ext (t), m → _ext ( t) It is good to have an estimator which derives the evaluation value of]. The sensor system may comprise one or more force sensors and / or torque sensors. The estimator may include a processor that executes a program for deriving the evaluation value of the craft window F → _ext (t).

であり、ここで、f→_D(t)は目標印加力であり、m→_D(t)は目標印加トルクであり、u_m,R1(t)は前記第１制御装置R1の制御信号成分であり、u_m,R2(t)は前記第２制御装置R2の制御信号成分である。ｔは時間である。また、前記目標クラフトヴィンダ−F→_D(t)は、前記ロボットに与えられたタスクに基づいて導出されるものである。 The device proposed here further comprises a control device connected to the first unit and the plurality of actuators AKT _m . The control device includes a first control device R1 which is an application force control device and a second control device R2 connected to the first control device R1, the second control device R2 being an impedance control device, It is an admittance controller, a position controller, or a speed controller. The control device generates a plurality of control signals u _m (t), and the plurality of actuators AKT _m are controlled by the plurality of control signals u _m (t) to operate the end effector When the end effector is in contact with the surface of the object, the surface is represented by the target craft Vinda-F → _D (t) = [f → _D (t), m → _D (t)] The target application force / target application torque is applied,
In the above,

Where f → _D (t) is the target application force, m → _D (t) is the target application torque, and um _{, R1} (t) are control signal components of the first control device R1. And um _{, R2} (t) are control signal components of the second control device R2. t is time. In addition, the target craft window F → _D (t) is derived based on a task given to the robot.

であり、ここで、u_m,R1 ^*(t)は目標クラフトヴィンダ−F→_D(t)で表される目標印加力／目標印加トルクを発生させるために前記第１制御装置R1が生成する制御信号であり、R→(t)は前記制御装置が受取る制御偏差であり、S(v(t))はF→_D(t)及びR→(t)に応じて値が定まるv(t)の単調減少関数であり、S^*(v^*(t), u_m,R1(t)^*)はu_m,R1(t)の影響度が基本的にＱ個の成分[v₁ ^*(t), v₂ ^*(t), ..., v_Q ^*(t)]の各々において単調減少する汎関数であり、[v_a, v_e]は変数v(t)の指定された定義域であり、[v_1a, v_1e], [v_2a, v_2e], ..., [v_Qa, v_Qe]はＱ次元のベクトル量v^*(t)の各成分の指定された定義域である。 Furthermore, the first control device R1 performs the control signal component um _{, R1} (t) as a product of the control signal um _{, R1} (t) ^* and a function S (v (t)), or a functional It is configured to generate as S ^* (v ^* (t), u _{m, R 1} (t) ^* ),
In the above,

, Where u _{m, R1} ^* (t) is generated by the first control device R1 to generate a target applied force / target applied torque represented by a target crafted vinder F → _D (t) Control signal, R → (t) is a control deviation received by the controller, and S (v (t)) is a value v ( _D ) determined according to F → _D (t) and R → (t). S ^* (v ^* (t), um _{, R1} (t) ^* ) is a monotonically decreasing function of t), and the influence of _{um, R1} (t) is basically Q components [v ₁ ^{* _{(t), v 2 * (}} t), ..., v Q * (t)] is a monotonically decreasing functionals in each, [v _a, v _e] was specified variable v (t) The domain, [v _1a , v _1e ], [v _2a , v _2e ], ..., [v _Qa , v _Qe ] is the specified component of the Q-dimensional vector quantity v ^* (t) It is a domain.

According to one configuration example of the present invention, the control signal component um _{, R1} (t) of the application force control device R1 is further reduced by a monotonically decreasing function S (v (t) to that generated by the conventional method. ) = S (v (F → _D (t), R → (t)) (a so-called shaping function), so that when the contact between the end effector and the surrounding environment disappears In one further configuration example, the control signal component u _{m, R1} (t (t) is used to generate the “shaping” effect. ) Is generated as a functional S ^* (v ^* (t), u _{m, R1} (t) ^* ) Furthermore, according to the present invention, the applied force control device has any structure. and can also generate a "shaping" effect. Moreover, the table as the product of the mathematically, the control signal u _{m, R1} (t) ^* a function S (v (t)) It is also possible to generate a can not "Shaping" effect of being. Therefore, as an example, monotonic decrease differed respectively for each control element of the PID controller shaping function ^{_{S * 1 (v * 1 (}} t)) , S ^* ₂ (v ^* 2 (t)), ... may be applied.

The function S (v (t)) has its value range set to [1, 0], and when the end effector and the surrounding environment are in contact with each other (ie, in the normal drive state), It is preferable that S (v (t)) = 1. If the contact between the end effector and the surrounding environment disappears, at the same time, the control deviation R → (t) received by the controller increases. The function S (v (t)) represents a target applied force / target applied torque to be applied from the end effector as the control deviation R → (t) increases, and a target craft window F → _D (t It is preferable that the value of this function S (v (t)) decreases from “1” to “0” more quickly as This is also true for the function S ^* (v ^* (t)).

In one configuration example of the device proposed here, the operation target (the operation target is in contact with the end effector, and the target application force represented by the target craft winder F → _D (t) (The object to which the target application torque is applied) has elastic deformability, and hence the control signal u _m (t) when the surface of the operation object has flexibility. The given elastic properties of the object in question are taken into account by the control device in the generation of.

One configuration of the device proposed here comprises a second unit, which serves as an energy reservoir for passivating the control device, given energy It is a unit for storing energy T1 to be delivered from the control device and supplying energy T2 to the control device according to the reservoir dynamics, and a closed loop control circuit by the second unit and the control device There is configured, the robot Manipyure - the amount of energy which data is corresponding to the energy consumption value E _Aufwand is derived or given value representing the amount of energy consumed in executing the actual task T0 The initial energy storage amount setting of the second unit is performed so as to be stored. Also in this regard, the energy E stored in the second unit may be virtual energy or physical energy. In the former case, the virtual energy is a mere calculated value (operand). In the latter case, the energy is some physical energy (e.g. power) and the second unit is a physical energy reservoir (e.g. battery) corresponding to the energy. Also, in the latter case, the control of the robot manipulator is improved, that is, not only passivation comes to be applied, but additionally, when the robot manipulator is driven. Energy consumption can also be reduced.

  In the configuration example described above, the energy upper limit value G1 is determined, and the second unit is configured such that the energy E stored in the second unit always becomes E ≦ G1. Is preferred. Furthermore, according to the energy E stored in the second unit, when the energy lower limit value G2 is set such that 0 <G2 <G1 and G2 <E ≦ G1, the second condition is satisfied. The second unit may be configured such that the second unit is disconnected from the control device when a unit is connected to the control device and E ≦ G2. preferable.

Another aspect of the present invention relates to a robot comprising a robot manipulator, the robot manipulator comprising an end effector, and a plurality (M) of actuators AKT _m (here: , M = 1, 2,..., M) and is characterized in that it comprises the device described above.

であり、ここで、f→_D(t)は目標印加力であり、m→_D(t)は目標印加トルクであり、u_m,R1(t)は前記第１制御装置R1の制御信号成分であり、u_m,R2(t)は前記第２制御装置R2の制御信号成分であり、以上において、前記第１制御装置R1は、前記制御信号成分u_m,R1(t)を、制御信号u_m,R1(t)^*と関数S(v(t))との積として、または、汎関数S^*(v^*(t), u_m,R1(t)^*)として生成するように構成されており、
以上において、

であり、ここで、u_m,R1 ^*(t)は目標クラフトヴィンダ−F→_D(t)で表される目標印加力／目標印加トルクを発生させるために前記第１制御装置R1が生成する制御信号であり、R→(t)は前記制御装置が受取る制御偏差であり、S(v(t))はF→_D(t)及びR→(t)に応じて値が定まるv(t)の単調減少関数であり、S^*(v^*(t), u_m,R1(t)^*)はu_m,R1(t)の影響度が基本的にＱ個の成分[v₁ ^*(t), v₂ ^*(t), ..., v_Q ^*(t)]の各々において単調減少する汎関数であり、[v_a, v_e]は変数v(t)の指定された定義域であり、[v_1a, v_1e], [v_2a, v_2e], ..., [v_Qa, v_Qe]はＱ次元のベクトル量v^*(t)の各成分の指定された定義域である。 Another aspect of the present invention is a robot having an end effector and driven by a plurality (M) of actuators AKT _m (where m = 1, 2,..., M). The present invention relates to a method of controlling a manipulator, which is represented by a Kraft-Vinder-F → _ext (t) = [f → _ext (t), m, which represents the external force / external torque acting on the end effector. → detects and / or derives _ext (t), where f → _ext (t) represents an external force acting on the end effector, and m → _ext (t) acts on the end effector It represents the external torque. The method further comprises the step of generating a plurality of control signals u _m (t) by means of a controller, said controller being connected to a first controller R1 which is an applied force controller and to the first controller R1. And a second control device R2, which is an impedance control device, an admittance control device, a position control device, or a speed control device, and the plurality of actuators AKT _m Under the control of a plurality of control signals u _m (t), when the end effector is in contact with the surface of the operation target, the end effector is placed on the surface, and the target craft vinder F → _D ( t) The target applied force / target applied torque represented by [f → _D (t), m → _D (t)] is applied,
In the above,

Where f → _D (t) is the target application force, m → _D (t) is the target application torque, and um _{, R1} (t) are control signal components of the first control device R1. Um _{, R2} (t) is a control signal component of the second control device R2, and in the above, the first control device R1 controls the control signal component um _{, R1} (t) Configured to generate as a product of u _{m, R1} (t) ^* and the function S (v (t)) or as a functional S ^* (v ^* (t), u _{m, R1} (t) ^* ) Has been
In the above,

, Where u _{m, R1} ^* (t) is generated by the first control device R1 to generate a target applied force / target applied torque represented by a target crafted vinder F → _D (t) Control signal, R → (t) is a control deviation received by the controller, and S (v (t)) is a value v ( _D ) determined according to F → _D (t) and R → (t). S ^* (v ^* (t), um _{, R1} (t) ^* ) is a monotonically decreasing function of t), and the influence of _{um, R1} (t) is basically Q components [v ₁ ^{* _{(t), v 2 * (}} t), ..., v Q * (t)] is a monotonically decreasing functionals in each, [v _a, v _e] was specified variable v (t) The domain, [v _1a , v _1e ], [v _2a , v _2e ], ..., [v _Qa , v _Qe ] is the specified component of the Q-dimensional vector quantity v ^* (t) It is a domain.

In the above method, when the operation object has elastic deformability, and the surface of the operation object has flexibility, therefore, the operation object is generated when the control signal u _m (t) is generated. Preferably, the given elastic properties of the object are taken into account by the control device.

Still further, a second unit is provided, said second unit functioning as an energy reservoir for passivating said control device, said control device according to a given energy reservoir dynamics A unit that stores energy T1 to be delivered from the unit and supplies energy T2 to the control device, and the second unit and the control device constitute a closed loop control circuit, the robot The second unit such that it stores energy T0 of an amount according to an energy consumption value E _Aufwand which is a derived or given value representing the amount of energy consumed by the manipulator performing an actual task Preferably, the initial energy storage setting is performed.

  By providing the embodiments described above with the same or corresponding features as given to the previously proposed device, various advantageous and preferred configurations of the proposed method can be obtained.

  Another aspect of the present invention relates to a computer system provided with a data processing apparatus, said data processing apparatus comprising: Data processing apparatus configured to be executed on the data processing apparatus.

  Another aspect of the present invention relates to a digital recording medium on which an electronically readable control signal is recorded, in which the control signal is cooperable with a programmable computer system. In operation, the method performed as described above is characterized in that it is performed.

  Another aspect of the present invention relates to a computer program product comprising a program code recorded on a machine-readable medium, said computer program product comprising The program code is executed on the data processing apparatus, whereby the method is executed as described above.

  Another aspect of the present invention relates to a computer program comprising program code, which is executed on a data processing apparatus. Is characterized in that the method is carried out as described above.

Therefore, the device proposed here as well as the method proposed here comprises an end effector and a plurality (M) of actuators AKT _m (where m = 1, 2,..., M) Apparatus and method for controlling a driven robot manipulator, wherein the application force control unit is combined with an impedance control unit (see FIG. 1), and further by combining an energy storage device It is based on the idea that the system is robust, while being based on According to the present invention, the surrounding environment, which is a passive element, may be anything, and it is not necessary to adjust the change of the virtual target position, which tends to lack robustness in its characteristics. According to the present invention, the manipulation of the surrounding environment by the robot manipulator can be made robust, compliant, and stable, and between the applied force control and the impedance control. There is no need to make a selection at. Furthermore, the disadvantages inherent in the applied force control and the impedance control described above are eliminated, and the advantages of the applied force control and the advantages of the impedance control are combined in the best possible manner. Become. Also, in particular, the dangerous movements of the robot manipulators which have hitherto occurred when the abutment between the end effector and the surrounding environment has disappeared.

  The following description of the embodiments will explain the invention in detail with respect to the following themes. A) Modeling of robot B) Basic configuration of control device C) Stabilization against loss of contact state D) Operation of operation target with flexibility and high compliance

ここで、q ∈ Rⁿは関節の回転角ポジションである。M(q) ∈ R^n×nは質量行列であり、C(q, q^・)q^・ ∈ Rⁿはコリオリ力及び遠心力を表すベクトルであり、g(q) ∈ Rⁿでは重力を表すベクトルである。この系の制御入力は、モ−タの駆動トルクτ_m ∈ Rⁿであり、また、τ_ext ∈ Rⁿには外部から作用する全ての外トルクが含まれる。説明を簡明にするために、ここでは摩擦力を無視している。外部からの作用力は、座標空間におけるベクトルF_ext ＝ (f_ext ^T, m_ext ^T)^T ∈ R⁶で与えられ、このベクトルは外力及び外トルクを表すものである。また、このベクトルは、転置ヤコビ行列J^T(q) によって、外部から夫々の関節に作用する外トルクへと転写され、その転写はτ_ext ＝ J^T(q)F_extで表される。 A Modeling of Robot
A1 The dynamics of a rigid robot manipulator having n rigid body dynamics joints (i.e., with n degrees of freedom (DOF)) is known and is given by the following equation (5).

Here, q ∈ R ⁿ is the rotational angle position of the joint. ^{M (q) ∈ R n ×} n is the mass matrix, C (q, q ^·) is q ^· ∈ R ⁿ is a vector representing the Coriolis force and centrifugal force, representing the g (q) ∈ gravity in R ⁿ It is a vector. The control input of this system is the drive torque τ _m τ R ⁿ of the motor, and τ _ext ∈ R ⁿ includes all external torques acting from the outside. In order to simplify the explanation, the friction force is ignored here. The external acting force is given by the vector F _ext = (f _ext ^T , m _ext ^T ) ^T ∈ R ⁶ in the coordinate space, and this vector represents external force and external torque. Also, this vector is transferred by the transposed Jacobian matrix J ^T (q) from the outside to an external torque acting on each joint, and the transfer is expressed by τ _ext = J ^T (q) F _ext .

ここで、θ ∈ Rⁿはモ−タの回転角ポジションである。式（6）は出力側のダイナミクスを記述した式であり、式（7）は入力側のダイナミクスを記述した式である。式（8）は関節トルクτ_J ∈ Rⁿによって式（6）と（7）を関連付けた式であり、この関節トルクは、線形バネ特性を有するバネにより得られている。当業者であれば、この式（8）をバネが非線形バネ特性を有する場合にまで拡張することも容易であろう。ここでは関節における減衰を無視することにする。減衰を考慮するように拡張することは容易であるため、それについてはここで論じない。行列K ∈ R^n×n及び行列B ∈ R^n×nはいずれも定数正定値対角行列であり、前者は関節の剛性を記述した行列、後者はモ−タの慣性を記述した行列である。また、ここでは更に、入力側の摩擦及び出力側の摩擦も考慮外としている。 A2 Dynamics of flexible joints In the case of a robot manipulator having a lightweight structure or a robot manipulator in which a spring is incorporated in a joint, the above-mentioned formula (5) has a structure having such flexibility. It is not possible to describe with sufficient accuracy the dynamics that inevitably occur during driving because of the existence of. Therefore, for a robot manipulator of such a structure, a model of a robot manipulator in which the joints are made resiliently rotatable (i.e., corrected as such) is used. This model is represented by the following formulas (6) to (8) (see Non-Patent Document 13 (Spong (1987))).

Here, θ ∈ R ⁿ is the rotational angle position of the motor. The equation (6) describes the dynamics on the output side, and the equation (7) describes the dynamics on the input side. The equation (8) relates the equations (6) and (7) by the joint torque τ _J関節 R ⁿ , and the joint torque is obtained by a spring having a linear spring characteristic. Those skilled in the art will easily extend this equation (8) to the case where the spring has non-linear spring characteristics. We will ignore the damping at the joints here. As it is easy to extend to take attenuation into account, it will not be discussed here. The matrix K ∈ R ^{n × n} and the matrix B ∈ R ^{n × n} are both constant positive definite diagonal matrices, the former is a matrix describing joint stiffness, and the latter is a matrix describing motor inertia. . Also, here, the friction on the input side and the friction on the output side are not taken into consideration.

Basic configuration of B controller
B1 Impedence control of a robot manipulator having a joint capable of elastic rotation and deflection by providing impediment control of the impedance control system of the B coordinate system. can do. In this passivation, for example, in order to feed back the position, it is preferable to feed back the value of the function of θ instead of feeding back the values of θ and q. In that case the value its static etc. q q ^- (θ) = ζ to be replaced by ^-1 (theta), the static equivalent value is implicit function _{ζ (q θ) = q θ} + K -1 g ( It calculated by contraction at q _theta), where qθ is the position of the output side of the equilibrium point. Under conditions not very strict, q ^- can be used (theta) as an estimate of q. For a more detailed explanation of the implicit function ζ, and the theory underlying it, see Albu-Schaffer, Ott, & Hirzinger, A Unified Passivity-based Control Framework for Position, Torque and Impedance Control of Flexible. See Joint Robots, (2007) The control equation for immobilizing impedance control of a robot manipulator with elastically rotationally displaceable joints is the following equation (9) And (10).

ここで、行列K_d ∈ R^6×6及び行列K_i ∈ R^6×6はいずれも正定値対角行列であり、前者は微分制御部分を表しており、後者は積分制御部分を表している。I ∈ R^6×6は単位行列であり、行列K_p ∈ R^6×6は、Kp − Iもまた正定値対角行列となるように選択されている。周囲環境へ印加しようとする目標印加力F_d :＝ (f_d ^T, m_d ^T)^Tは、作業者であるオペレ−タもしくは作業計画の作成者によって指定される。尚、以下の説明では、式を見やすくするために、h_i(F_ext(t), t) :＝ K_i∫₀ ^t(F_ext(t) − F_d(t))dσと書き替えることにする。F_extの値は、力センサから得るようにすることもでき、モニタ装置から得るようにすることもできる（非特許文献７（Haddadin, Towards Safe Robots; Approaching Asimov's 1st Law（2013年）参照）。この印加力制御装置が用いられるロボット・マニピュレ−タが、剛体ロボット・マニピュレ−タである場合には、上式（5）においてτ_m ＝ τ_mfとし、また、それが弾性的回転偏位可能な関節を備えたロボット・マニピュレ−タである場合には、上式（7）においてτ_m ＝ τ_mfとすればよい。 The basic configuration of the application force control device controller of the B2 coordinate system is based on the application force control device of the coordinate system whose control equation is expressed by the following equation (11).

Here, the matrix K _d ∈ R ^{6 × 6} and the matrix K _i ∈ R ^{6 × 6} are all positive definite diagonal matrices, the former representing a differential control part and the latter representing an integral control part . I ∈ R ^{6 × 6} is an identity matrix, and the matrix K _p ∈ R ^{6 × 6} is chosen such that Kp − I is also a positive definite diagonal matrix. The target application force F _d : = (f _d ^T , m _d ^T ) ^T to be applied to the surrounding environment is specified by the operator who is the operator or the creator of the work plan. In the following description, in order to make the expression easy to read, rewrite h _i (F _ext (t), t): = K _i ∫ ₀ ^t (F _ext (t) − F _d (t)) dσ Make it The value of F _ext can be obtained from a force sensor or from a monitoring device (see Non-Patent Document 7 (Haddadin, Towards Safe Robots; Approaching Asimov's 1st Law (2013)). If the robot manipulator in which this applied force control device is used is a rigid robot manipulator, then τ _m = τ _mf in the above equation (5), and it is possible to make an elastic rotational deflection. In the case of a robot manipulator having a flexible joint, it is sufficient to set τ _m = τ _mf in the above equation (7).

B3 Integrated control device integrating the application force control device and the impedance control device The control equation in the case where the application force control device described above and the above-described impedance control device are simply combined is the following equation (12 It becomes what was shown to).

ここで、x_tはエネルギ貯留器の状態を表すものであり、また、ωは下式（14）によって定義される。

However, in such a case, stability is not guaranteed. Therefore, the control device so configured needs to be discussed in conjunction with the energy reservoir (see FIG. 2) to ensure that the energy reservoir performs system passivation and thus the system To ensure the stability of the The control equation in such a case is as shown in the following equation (13).

Here, x _t represents the state of the energy storage, and ω is defined by the following equation (14).

ここで、u_τはエネルギ貯留器への入力を表すものである。また、バイナリのスカラ−値であるα、s、γによって、系全体の安定性が常時保証されている。 The dynamics of the energy reservoir can be described as equation (15) below.

Here, u _τ represents the input to the energy reservoir. Moreover, the stability of the whole system is always guaranteed by the binary scalar values α, s and γ.

B4 Initial setting of energy storage amount for task base Calculation of required energy (energy required to execute task) is based on static applied force equivalent weight f _I | _{x = xw} + f _d = f _w Where f _I = K _{x, t} (p-p _s ) represents the force on the stiffness of the impedance and f _w = K _{w, t} (p _w -p _{w, 0} ) is It represents the force against the reaction force from the surrounding environment. The reaction force is generated on the surface of the operation target, and here, the reaction force is modeled in the form of a linear function taking into consideration only the rigidity without considering the damping. By solving for p _w , the position of the surface is obtained, and the control of the applied force is performed according to the position thus obtained. Therefore, the amount of work required to move the surface (i.e., the energy required for task execution) is calculated by the following equation (16).

  However, the above equation (16) considers only the energy required for translational movement. In the case of rotational movement, of course, it is necessary to extend the same equation, but the extension can be easily performed by those skilled in the art.

As a special control form, control may be performed such that f _d = const., And the task execution required energy in that case is calculated by the following equation (17).

  Then, initial energy storage amount setting is performed so as to store the task execution required energy.

C Stabilization against loss of contact The robot manipulator is guaranteed to be stable in its control under all conceivable conditions, but it is guaranteed that the stability of its control is guaranteed. Manipulators do not immediately mean that only safe exercise occurs. When the contact between the end effector and the surface of the operation object unexpectedly disappears, the robot equipped with the robot manipulator recovers the applied force that has disappeared due to the disappearance of the contact. It may happen that the operation to be performed is continued until the energy stored in the energy reservoir is zero. Depending on the amount of energy stored in the energy storage, the movement of the robot manipulator thus generated is very large, the movement speed is also large, and it is possible for the movement to be most serious and inconvenient.

  As a method of avoiding such a situation, it is not possible to simply stop the control device immediately if the contact state between the end effector and the surrounding environment or the operation target is not detected. is not. However, when such a method is used, for example, when noise is mixed in the sensor output, there is a possibility that an undesirable on / off operation may occur.

この関数は、並進移動の部分と回転移動の部分とを含んでおり、それら部分は夫々以下のように定義されている。

及び

以上に具体例として示した関数ρ(ψ)は、上述の関数 S(V) に対応しており、また、変数ψは、上述の制御偏差R→(t)に対応している。 In order to address this problem, here we propose a robust position-based control method. This method is performed using a controller shaping function S (V): == (ψ), which is integrated into the controller. This function is expressed by the following equation.

This function includes a part of translational movement and a part of rotational movement, which are respectively defined as follows.

as well as

The function ρ (ψ) described above as a specific example corresponds to the function S (V) described above, and the variable 対 応 corresponds to the control deviation R → (t) described above.

The position x of the robot manipulator x: = (p ^T , φ ^T ) ^T includes a translational movement portion (translational movement amount) p and a rotational movement portion (rotational amount) φ, and the rotation amount φ Is represented by an appropriate rotation angle display method such as, for example, Euler-corner. Here, Δ p = p _s − p is a vector starting from the end effector and ending at the virtual target position, F _d = (f _d ^T , m _d ^T ) ^T is a six-dimensional target craft window and (See Figure 2). Then, if the angle formed by the Δp and f _d becomes larger than 90 °, so that that the operation stop state controller at that time.

An interpolation processing function _{t t} (ψ for performing interpolation processing in a defined region d _max designated by the operator so that state transition between the operating state and the operation stopping state is smoothly performed. ) Is selected. Further, as a method of displaying the rotation amount _{r r} (ψ), a quaternion which is an angle display method of a singular point is selected. The unit quaternion k = (k ₀ , k _v ) represents the real direction, and the quaternion K _s = (k _{0, s} , k _{v, s} ) represents the target direction.

Therefore, the rotation angle deviation Δk: = k ^-1 k _s and [Delta] [phi: = 2 represented by arccos (Δk _0). Further, a defined area designated as a robust area by the operator is indicated by a rotation angle φ _max . The angle of rotation speed of the quaternion scalar - is associated with the _{component, φ max: = 2 arccos (} k 0, max) become. From the viewpoint of stability analysis, the shaping function is considered to be a function that applies shaping to ω, which is one of integrated control devices in which an applied force control device and an impedance control device are integrated. The part of the application force control device is nothing but performing a scaling process which is multiplied by a scaling factor. By this processing, ωφ: = ρ (ψ) ω used in place of ω is obtained, and stability is thereby guaranteed. Here, the process of multiplying ρ (ψ) is performed on each component of ω.

D As described in the previous chapter on the operation of the operation target with flexibility and high compliance, when performing control by the position-based control method, for a material that is flexible and easy to be deformed It needs to be considered that special handling is required. When designing the virtual target position, if it is designed without considering the compliance or deformation of the surrounding environment, there is a problem that the application force control device may cause an ineffective stop or scaling depending on the situation. It's no good. These adverse events occur because the presence of compliance causes the actual position to be a different position than in the absence of compliance. Therefore, if the material of the object to be manipulated is flexible and the compliance is high, the corrected virtual target position x _d ′ = (p, in order to match the virtual target position of the controller with such material. _d ' ^T , φ _d ' ^T ) Introduce ^T. Here, if K _mat represents a determination value (not necessarily a known value) of the stiffness value of the surface or object to which an operation is to be applied, the correction equation for quasi-static correction is 18).

The above equation (18) can also be abbreviated as x _d ′ = x _d −K _mat ⁻¹ F _d . As apparent from this equation, here, the virtual target position is shifted so as to offset the deflection generated in the material having flexibility or elastic deformation (see FIG. 4). Of course, in applying this method, the stiffness value of the surrounding environment must be obtained as a known value or at least as an estimated value. However, the situation described above when K → さ れ る is intuitively understood from the above equation (18). Also, as a matter of course, the above equation (18) can be extended to an equation taking into account the attenuation of the surrounding environment. However, if such expansion is performed, the amount of arithmetic processing increases.

  Various other advantages, features, and details are set forth in, and will be apparent from, the following detailed description, in conjunction with at least one embodiment, which will be described with occasional reference to the drawings. In the drawings, the same components, the same components, and / or the components having the same functions are denoted by the same reference numerals.

ac is a schematic conceptual diagram for demonstrating the hybrid control apparatus which integrated an impedance control apparatus and an application force control apparatus which is proposed here. FIG. 6 is a control system diagram showing a specific example of a control system that controls robot manipulators that interact with the surrounding environment. FIG. 16 is a schematic view showing a robot manipulator provided with an end effector EFF, to which impedance control is applied, and in which a robust region d _max for translational movement is designated. It is the figure which showed the control device shaping function (phi) ((zeta)) = S (v (t)) used when performing translational movement. FIG. 10 is a schematic view showing a deformation that occurs in a material having elastic deformability by an applied force applied from an end effector when performing translational movement. FIG. 1 is a schematic structural view of an apparatus proposed herein. It is a flowchart which showed the implementation plan of the method proposed here.

FIGS. 1 a to 1 c are schematic conceptual diagrams for explaining a hybrid control device integrated with an impedance control device and an application force control device as proposed herein. Attenuation elements are not shown for the sake of clarity. The robot manipulator provided with the end effector EEF shown in FIG. 1a is applied only to impedance control. The fact that impedance control is applied is represented by a spring element in the figure. The robot manipulator equipped with the end effector EFF shown in FIG. 1b is applied with only applied force control, and the end effector EFF is shown by the operation target object surface (hatched line on one side The target application force F _d is applied to FIG. 1c shows a control system according to the invention in which the impedance control system shown in FIG. 1a and the application force control system shown in FIG. 1b are integrated.

  FIG. 2 is a control system chart showing a specific example of a control system for controlling a robot manipulator having an end effector according to the present invention, and this robot manipulator is an ambient environment / operation object / Workpieces, etc. interact with each other. Several elements are shown in the form of functional blocks in the figure, and the surrounding environment (functional block noted as "ambient environment") is a robot manipulator (functional block noted as "rigid body dynamics") and plural It interacts with each other (blocks marked “Motor Dynamics”). The control of the plurality of actuators is performed by a control unit (a functional block described as “application force / impedance control unit”), and in this control unit, an energy storage unit (“energy storage unit”) is described. Functional blocks are connected in a connectable and disconnectable manner. The connection relationship between these interconnected functional blocks and the feedback path are shown by the input / outputs related to them and the amount of input / output transferred between the input / outputs. . The feedback of the external force disappears if the controller and the energy reservoir are disconnected so that the operation of the controller is not damaged.

FIG. 3a is a schematic diagram showing a robot manipulator comprising an end effector EFF, to which impedance control is applied, and in which a robust region d _max for translational movement is designated. Δ p = p _s − p is a vector starting from the position p of the end effector and ending at the operation action point p _s , and f _b is a target craft window.

  FIG. 3 b is a diagram showing a controller shaping function φ (ψ) = S (v (t)) used when performing translational movement. A detailed description of an embodiment of this function ψ (ψ) is given above (see chapter C "Stabilization against loss of contact").

  FIG. 4 is a schematic view showing a deformation that occurs in a material having elastic deformability by an applied force from the end effector EFF when performing translational movement, and the above description (Chapter D As described above in detail, the operation of the operation object having flexibility and high compliance.

FIG. 5 shows a control device as proposed here, namely a robot manipulator comprising an end effector and driven by three actuators AKT _m (where m = 1, 2, 3) FIG. 2 is a schematic structural view of an apparatus for This device detects and / or derives a craft vinder F → _ext (t) = [f → _ext (t), m → _ext (t)] representing external force / external torque acting on the end effector The first unit 101 is provided, where f → _ext (t) represents an external force acting on the end effector, and m → _ext (t) represents an external torque acting on the end effector. The apparatus further comprises a control unit 102 connected to the first unit 101 and the actuator AKT _m , the control unit 102 comprising: a first control unit R1 which is an application force control unit; And a second controller R2 which is an impedance controller connected to the device R1. The controller 102 generates a plurality of control signals u _m (t), and the actuator AKT _m is controlled by the plurality of control signals u _m (t) so that the end effector contacts the surface of the operation target. When in contact with the surface, the target effecter / target represented by the end effector on the surface is represented by: Target Craft V inder F → _D (t) = [f → _D (t), m → _D (t)] An applied torque is applied. In the above, u _m (t) = u _{m, R 1} (t) + u _{m, R 2} (t), where f → _D (t) is the target application force and m → _D (t) is The target applied torque is um _{, R1} (t) is a control signal component of the first control device R1, and um _{, R2} (t) is a control signal component of the second control device R2. Also, in the above, the first control device R1 is configured to generate the control signal component u _{m, R1} (t) as the product of the control signal u _{m, R1} (t) ^* and the function S (v (t)) In the above, um _{, R1} (t) = S (v (t)) um _{, R1} (t) ^* , and v (t) = v (F-> _D (t), R-( a t)), v a _{(t) ∈ [v a,} v e], further wherein the target u _{m, R1} ^* (t) is represented by the target craft Vin d'-F → _D (t) A control signal generated by the first controller R1 to generate an applied force / target applied torque, R → (t) is a control deviation received by the controller 102, and S (v (t)) is F → a monotonically decreasing function of _D (t) and R → (t) value is determined according to v (t), a [v _a, v _e] the domain that is specified in the variable v (t).

FIG. 6 shows the control method proposed here, ie, an end effector, which is driven by a plurality (M) of actuators AKT _m (where m = 1, 2,..., M) Is a flowchart showing an implementation plan of a method of controlling a robot manipulator. The method comprises the following steps: In the first step 201, a craft vinder F → _ext (t) = [f → _ext (t), m → _ext (t)] representing external force / external torque acting on the end effector is detected and / or Deriving, where f → _ext (t) represents an external force acting on the end effector, and m → _ext (t) represents an external torque acting on the end effector. In the second step 202, a plurality of control signals u _m (t) are generated by the control unit 102, and the control unit 102 is connected to a first control unit R1 which is an application force control unit and the first control unit R1. And a second controller R2 which is an impedance controller. A plurality of actuators AKT _m are controlled by a plurality of control signals u _m (t) so that when the end effector is in contact with the surface of the operation target, the end effector is on the target surface, the target A target application force / target application torque represented by the following equation is applied: Kraft vinder F → _D (t) = [f → _D (t), m → _D (t)]. In the above, u _m (t) = u _{m, R 1} (t) + u _{m, R 2} (t), where f → _D (t) is the target application force and m → _D (t) is The target applied torque is um _{, R1} (t) is a control signal component of the first control device R1, and um _{, R2} (t) is a control signal component of the second control device R2. Furthermore, in the above, the first control device R1 sets the control signal component um _{, R1} (t) as a product of the control signal um _{, R1} (t) ^* and the function S (v (t)), or a functional S ^* (v ^* (t), um _{, R1} (t) ^* ) is generated, and in the above, um _{, R1} (t) = S (v (t)) um _{, R1} (t) ^* , v (t) = v ( F → D (t), R → (t)) is, v (t) ∈ [v a, v e] is further _wherein, u _{m, R1} ^* (t ) Is a control signal generated by the first control device R1 to generate a target application force / target application torque represented by a target craft type F → _D (t), where R → (t) is a control device 102 is the control deviation received, and S (v (t)) is a monotonically decreasing function of v (t) whose value is determined according to F → _D (t) and R → (t), [v _a , v _e ] is a designated domain of the variable v (t).

  Although the present invention has been illustrated and described in detail in connection with the preferred embodiments, the scope of the present invention is not limited to the embodiments disclosed above, and those skilled in the art can protect the rights of the present invention. It goes without saying that the present invention can be applied to other various configuration examples based on the embodiment without departing from the scope. That is, it is obvious that there can be many different configurations that can be realized. Furthermore, it is also apparent that the embodiments disclosed as the examples merely exemplify the specific examples, and the scope of protection of rights, uses, or configurations of the present invention are not limited at all. Rather, the foregoing detailed description and the description with reference to the drawings enable a person skilled in the art to make the illustrated embodiments a specific configuration, and those skilled in the art who have known the concept of the present invention disclosed above. However, various modifications related to, for example, the functions or configurations of the individual components mentioned in connection with the embodiments can be made without departing from the scope of protection of the present invention. The scope of protection of the invention includes the features described in the claims, and also includes the features that are legally equivalent to the features described in the claims, as outlined in the specification. It is defined as a range.

Claims (14)

In an apparatus for controlling a robot manipulator comprising an end effector and driven by a plurality (M) of actuators AKT _m (where m = 1, 2,..., M),
_-A first method of detecting and / or deriving a craft vinder F → _ext (t) = [f → _ext (t), m → _ext (t)] representing an external force / external torque acting on the end effector Equipped with a unit (101)
here,
f → _ext (t) represents an external force acting on the end effector,
m → _ext (t) represents the external torque acting on the end effector,
A control unit (102) connected to the first unit (101) and the plurality of actuators AKT _m , the control unit (102) being a first control unit R1 being an applied force control unit and the control unit (102); A second controller R2 connected to the first controller R1, the second controller R2 being an impedance controller, an admittance controller, a position controller, or a speed controller, said control The device (102) generates a plurality of control signals u _m (t), and the plurality of actuators AKT _m are controlled by the plurality of control signals u _m (t) to operate the end effector When the end effector is in contact with the surface of the object, the end effector is represented by the target craft window F → _D (t) = [f → _D (t), m → _D (t)]. Target applied force / target applied torque is applied.
In the above,

And
here,
f → _D (t) is the target application force,
m → _D (t) is the target applied torque,
um _{, R1} (t) is a control signal component of the first control device R1,
um _{, R2} (t) is a control signal component of the second control device R2,
In the above,
The first control device R1 sets the control signal component um _{, R1} (t) as a product of the control signal um _{, R1} (t) ^* and a function S (v (t)) or a Q-dimensional pan. It is configured to generate as a function S ^* (v ^* (t), u _{m, R 1} (t) ^* ),
In the above,

And
here,
u _{m, R 1} (t) ^* is a control signal generated by the first control device R 1 to generate a target application force / target application torque represented by a target craft winder F → _D (t),
R → (t) is a control deviation received by the controller (102) ,
S (v (t)) is a monotonically decreasing function of v (t) whose value is determined according to F → _D (t) and R → (t),
S ^* (v ^* (t), um _{, R1} (t) ^* ) is a functional function in which the degree of influence of um _{, R1} (t) monotonously decreases,
[v _a , v _e ] is the designated domain of variable v (t),
[v _1a , v _1b ], ... is the designated domain of each component of the Q-dimensional variable v ^* (t),
An apparatus characterized by When the operation object has elastic deformability, and the surface of the operation object has flexibility, therefore, the control signal u _m (t) is generated when the operation object is given. Device according to claim 1, characterized in that the elastic properties of are taken into account by the control unit (102). A second unit, the second unit functioning as an energy reservoir for passivating the controller (102), the controller (102) according to a given energy reservoir dynamics And a unit for storing energy T1 to be delivered from and supplying energy T2 to the control device (102), and a closed loop control circuit is provided between the second unit and the control device (102). Energy consumption value E which is a derived or given value representing the amount of energy the robot manipulator consumes when performing an actual task, and storing energy T0 of an amount according to _Aufwand The apparatus according to claim 1 or 2, wherein an initial energy storage setting of the second unit is performed to do this.   The apparatus according to claim 3, wherein the energy E stored in the second unit is virtual energy or physical energy.   The energy upper limit value G1 is determined, and the second unit is configured such that the energy E stored in the second unit always satisfies E ≦ G1. Device described. The energy lower limit value G2 is determined to be 0 <G2 <G1, and according to the energy E stored in the second unit,
The second unit is connected to the controller (102) when G2 <E ≦ G1;
When E ≦ G2, the second unit is disconnected from the controller (102),
The apparatus of claim 5, wherein the second unit is configured. The first unit (101) detects a craft window F → _ext (t) = [f → _ext (t), m → _ext (t)] representing an external force / external torque acting on the end effector And / or evaluation of a craft vinder F → _ext (t) = [f → _ext (t), m → _ext (t)] representing the external force / external torque acting on the end effector 7. An apparatus according to any one of the preceding claims, characterized in that it comprises an estimator for deriving values. In a robot equipped with a robot manipulator, the robot manipulator comprises an end effector, and a plurality (M) of actuators AKT _m (where m = 1, 2, ..., M) An apparatus according to any one of claims 1 to 7 and driven by
A robot characterized by In a method of controlling a robot manipulator having an end effector and driven by a plurality (M) of actuators AKT _m (where m = 1, 2,..., M),
_-Detecting and / or deriving a craft vinder F → _ext (t) = [f → _ext (t), m → _ext (t)] representing the external force / external torque acting on the end effector 201),
here,
f → _ext (t) represents an external force acting on the end effector,
m → _ext (t) represents the external torque acting on the end effector,
A step (202) of generating a plurality of control signals u _m (t) by means of the control unit (102), said control unit (102) being a first control unit R1 being an applied force control unit and said first control unit A second controller R2 connected to R1, the second controller R2 being an impedance controller, an admittance controller, a position controller, or a speed controller, said plurality of actuators; The AKT _m is controlled by the plurality of control signals u _m (t) so that when the end effector is in contact with the surface of the operation target, the end effector is moved to the target craft vinder The target application force / target application torque represented by −F → _D (t) = [f → _D (t), m → _D (t)] is applied,
In the above,

And
here,
f → _D (t) is the target application force,
m → _D (t) is the target applied torque,
um _{, R1} (t) is a control signal component of the first control device R1,
um _{, R2} (t) is a control signal component of the second control device R2,
In the above,
The first control device R1 sets the control signal component u _{m, R1} (t) as a product of the control signal u _{m, R1} (t) ^* and the function S (v (t)), or the functional S ^* generated as (v ^* (t), u _{m, R 1} (t) ^* ),
In the above,

And
here,
u _{m, R 1} (t) ^* is a control signal generated by the first control device R 1 to generate a target application force / target application torque represented by a target craft winder F → _D (t),
R → (t) is a control deviation received by the controller (102) ,
S (v (t)) is a monotonically decreasing function of v (t) whose value is determined according to F → _D (t) and R → (t),
S ^* (v ^* (t), um _{, R1} (t) ^* ) is a functional function in which the degree of influence of um _{, R1} (t) monotonously decreases,
[v _a , v _e ] is the designated domain of variable v (t),
[v _1a , v _1b ], ... is the designated domain of each component of the Q-dimensional variable v ^* (t),
A method characterized by When the operation object has elastic deformability, and therefore the surface of the operation object has flexibility, the generation of the control signal u _m (t) gives a given operation object. The method according to claim 9, characterized in that the elastic properties of are taken into account by the control unit (102). A second unit, the second unit functioning as an energy reservoir for passivating the controller (102), the controller (102) according to a given energy reservoir dynamics And a unit for storing energy T1 to be delivered from and supplying energy T2 to the control device (102), and a closed loop control circuit is provided between the second unit and the control device (102). Energy consumption value E which is a derived or given value representing the amount of energy the robot manipulator consumes when performing an actual task, and storing energy T0 of an amount according to _Aufwand Method according to claim 9 or 10, characterized in that an initial energy storage setting of the second unit is performed. A computer system comprising a data processing apparatus for causing a robot manipulator to execute , the method according to any one of claims 9 to 11 being executed on the data processing apparatus A computer system characterized in that the data processing apparatus is configured.   12. A method according to any of claims 9 to 11 in a digital recording medium having an electronically readable control signal recorded thereon, said control signal cooperating with a programmable computer system. Digital recording medium characterized by   10. A computer program product comprising a program code recorded on a medium readable by a machine, said program code being executed on a data processing device. 12. A computer program product characterized in that the method according to any one of the claims 11 to 11 is carried out.

Images